《安訊士（Axis）：2025視頻監控領域AI應用狀況研究報告（英文版）（23頁）.pdf》由會員分享，可在線閱讀，更多相關《安訊士（Axis）：2025視頻監控領域AI應用狀況研究報告（英文版）（23頁）.pdf（23頁珍藏版）》請在三個皮匠報告上搜索。

1、Industry perspectives on the transformation of security,safety,business intelligence and operational efficiencyThe state of AI in video surveillanceREPORT|2025ContentsExecutive summary 3Introduction 4Methodology 5KEY FINDINGS Top trends impacting the future 79 insights:From the acceleration of cloud

2、 and edge AI to growing use of facial recognition 9Deep-dive:Responsible AI in video surveillance 15Conclusions 19Axis approach to AI 20 Appendix 21THE STATE OF AI IN VIDEO SURVEILLANCE|3Executive summaryAs AI continues to transform video surveillance,its potential to improve security and safety,ope

3、rational efficiency,and business intelligence is increasingly recognized.This report examines the key opportunities and challenges and expected future trends in these areas but also how responsible practices are becoming crucial for companies in relation to their use of AI.Drawing insights from qual

4、itative research as well as quantitative data sources,it outlines the possibilities and hurdles that AI poses.Themes such as the balance between cloud and edge computing,ethical considerations,customer expectations,and the need for collaboration emerge.Looking ahead,AI in security,safety,and beyond

5、will likely focus on generative AI,integrated capabilities,smarter AI systems,and deeper IoT integration,while also navigating complex ethical and privacy issues.Key insights identifiedAI adoption surges with differences between markets The transition to cloud and edge AI continues to accelerate Int

6、egration of diverse data generating insightsGrowing use of facial recognitionNew opportunities for business intelligence and operational efficiencyAI integration challenges bridging expectations and capabilitiesPartnership and collaboration to enable AI integrationNavigating responsible AI The road

7、ahead-future trends to watch:generative AI,smarter AI pipelines,AI and IoT convergence123456789Executive summaryTHE STATE OF AI IN VIDEO SURVEILLANCE|4IntroductionThe rise of AI has created new possibilities for how organizations monitor and secure their environments.From facial recognition to predi

8、ctive analytics,AI-driven systems have transitioned from cutting-edge innovations to essential tools for enhancing safety,security,for providing business intelligence,and enhancing operational efficiency.However,as AI evolves,so do the challenges associated with its deployment.THE STATE OF AI IN VID

9、EO SURVEILLANCE|AI adoption offers significant advantages,including greater efficiency,improved decision-making,and enhanced output quality,benefiting businesses and society alike.However,it also presents challenges around privacy,cybersecurity,and technology integration,emphasizing the need for tru

10、st,responsibility,and accountability.At Axis,we are leading this transformation,working closely with partners,end customers,and other stakeholders to shape a future where AI enables a smarter,safer world.IntroductionTHE STATE OF AI IN VIDEO SURVEILLANCE|5A very special thanks to:This report explores

11、 opportunities and challenges shaping AI adoption in safety,security,business intelligence and operational efficiency.It combines insights from eleven in-depth interviews with carefully selected experts from Axis global partner network,each demonstrating a proven track record and extensive expertise

12、 in the field of AI.Technology partners and system integrators from across the world were selected and interviewed during 2024.Each interview lasted approximately one hour and followed a semi-structured format.To supplement the interview findings,data was drawn from two global surveys conducted by A

13、xis in Q2 2024:the annual Axis sales channel survey and the Axis end customer survey.The Axis sales channel survey was sent to distributors and channel partners in 68 countries,generating about 4,900 responses.The Axis end customer survey was distributed to end customers in 64 countries,yielding aro

14、und 900 responses.MethodologyMethodologyTHE STATE OF AI IN VIDEO SURVEILLANCE|6KEY FINDINGSThe key findings are divided into three parts:top trends,9 insights and a deep-dive on responsible AI.THE STATE OF AI IN VIDEO SURVEILLANCE|KEY FINDINGSTHE STATE OF AI IN VIDEO SURVEILLANCE|7The top 3 trends i

15、mpacting end customers Cybersecurity,risk and privacy emerge as the top priority for end customers,with 61%of respondents ranking it as a critical area of impact.Analytics and actionable insights follow,identified by 36%of end customers as significant.Finally,AI and generative AI are recognized by 3

16、4%of respondents as transformative technologies,impacting the future of their business.Which of the following trends do you consider most significant for your business in the near future?Maximum 3 answersCybersecurity,risk and privacy:61%Analytics and actionable insights:36%AI and generative AI:34%o

17、f end customers identify AI and generative AI as game-changing trends that will significantly shape the near future of their businesses.34%Top trends impacting the futureThe Axis sales channel survey and Axis end customer survey asked partners and end customers to identify the top trends impacting t

18、he video surveillance industry and their respective businesses.Both groups channel partners and end customers highlighted cybersecurity,risk and privacy,AI and generative AI,analytics and actionable insights,as some of the most significant trends shaping their industry and operations.Its worth notin

19、g that several of these areas could be perceived to overlap and influence each other.This is confirmed by the in-depth interviews,where many of these topics are discussed together,highlighting their interconnected nature.123KEY FINDINGS Top trendsTHE STATE OF AI IN VIDEO SURVEILLANCE|8The top 3 tren

20、ds impacting partnersSurvey results reveal that partners recognize cybersecurity,risk,and privacy as well as AI and generative AI as the two most significant trends shaping the video surveillance industry,with both ranked by 62%of respondents.50%of partners also highlight analytics and actionable in

21、sights as a key trend.Which of the following trends do you think will be most important for our industry in the near future?Multiple answers possible123of senior executives and top management identify AI and generative AI as the most significant industry trend poised to impact their business in the

22、future.of partners rank AI and generative AI as one of the most significant trends impacting the industry in the future.66%62%AI and generative AI:62%Cybersecurity,risk and privacy:62%Analytics and actionable insights:50%Senior management among partners name AI and generative AI as the leading indus

23、try trendAxis sales channel survey reveals the top three industry trends identified by senior management,highlighting their top priorities shaping the future of the video surveillance industry.The survey gathered insights from over 600 representatives in top and senior management.AI and generative A

24、I lead the rankings with 66%of senior leaders recognizing these technologies as the most significant trend.Cybersecurity,risk,and privacy follow closely,noted by 61%of senior executives.Rounding out the top three,analytics and actionable insights were ranked by 56%of senior management as a key trend

25、.KEY FINDINGS Top trendsTHE STATE OF AI IN VIDEO SURVEILLANCE|99 insights:From the acceleration of cloud and edge AI to growing use of facial recognitionTo complement the survey data with in-depth insights,eleven carefully selected AI experts from technology partners and system integrators within Ax

26、is partner network were interviewed.THE STATE OF AI IN VIDEO SURVEILLANCE|KEY FINDINGS 9 insightsTHE STATE OF AI IN VIDEO SURVEILLANCE|10AI adoption surges with differences between marketsInterviewees unanimously reported a significant buzz around AI,with interest skyrocketing over the past few year

27、s.Customers are not only becoming more curious about AI but are also growing increasingly knowledgeable about its potential applications.Despite this progress,most agreed that we are still in the early days of AI adoption,with much room for growth and refinement.AI adoption appears to be more advanc

28、ed in certain markets,such as the US,Japan,Singapore,and certain parts of Europe,where customers are more actively seeking AI solutions to improve surveillance and operational efficiency.Experts attributed this to variations in market maturity,regulatory frameworks,and customer readiness to invest i

29、n and deploy AI solutions.1The interview findings reveal several thematic insights:We are betting on edge computing,but ultimately,many solutions will require a balance between edge and cloud.Partner,AmericasAI is a valuable feature but can be challenging to both understand and use.Partner,APACThe t

30、ransition to cloud and edge AI continues to accelerate The shift from traditional on-premise server systems to hybrid architectures continues,where cloud-based technology is combined with edge AI solutions.This development is driven by the need for easier scalability,faster processing,and improved b

31、andwidth usage.Edge AI plays a crucial role by enabling faster,local analytics with minimal latency that is essential for real-time responsiveness in critical situations.At the same time,cloud integration is becoming increasingly important,especially for managing large installations on multiple site

32、s.The hybrid model,which combines the immediate processing capabilities of edge AI on cameras with the scalability and easy access to systems in the cloud,is emerging as the preferred approach by many.This balance allows organizations to harness the strengths of both technologies.By moving AI proces

33、sing closer to the source,on edge devices like cameras,companies can reduce bandwidth consumption,enhance efficiency,and better support real-time applications such as security monitoring.As this trend continues,the hybrid approach is expected to continue to shape the future of AI in security,safety,

34、business intelligence and operational efficiency.2KEY FINDINGS 9 insightsTHE STATE OF AI IN VIDEO SURVEILLANCE|11Integration of diverse data generating insightsThe integration of diverse data sources for a more comprehensive analysis is emerging as a trend that could revolutionize safety and securit

35、y while elevating business intelligence and improving operational efficiency.Experts predict that integrating additional sensory data,such as audio and contextual environmental factors,to complement video data will enhance situational awareness,provide more actionable insights,and offer a more compr

36、ehensive understanding of events.Experts emphasize that combining multiple streams of data enables more accurate detection and prediction of potential threats.For instance,in emergency scenarios,pairing visual data with audio analysis can allow security teams to respond faster and more precisely.Thi

37、s richer,context-aware approach reflects how human operators process multiple inputs to make informed decisions,representing a significant step forward in improving safety,security,and operational efficiency.Growing use of facial recognitionFacial recognition has become widely adopted in many countr

38、ies,particularly in Asia,where it is commonly used for urban security,surveillance,and commercial applications.In contrast,adoption seems to have been slower in Europe,the US and countries like Australia and New Zealand due 34Looking further ahead,well see AI combining data from multiple sources for

39、 better insights.Multimodality as a concept has been there for quite some time,but I think it will accelerate.And that will enable more intelligent decision-making.Partner,APACWe see customers integrating AI with additional sensor data to provide richer,more actionable intelligence.And that is a who

40、le new game.Partner,Americasto stricter privacy regulations and public concerns about surveillance.However,the introduction of new regulations is helping to clarify how facial recognition should be applied by differentiating use cases based on risk and providing a clear framework for its responsible

41、 useInterviewees predict that facial recognition will continue to gain traction globally.Ethical considerations around its use remain a central focus,particularly in regions with strong privacy laws.Most interviewees highlight that there are effective ways to implement facial recognition responsibly

42、,including aligning with privacy regulations and maintaining transparency about how the technology works and should be applied.This transparency is seen as critical to building and maintaining public trust in the technology.New opportunities for business intelligence and operational efficiencyThe in

43、tegration of AI is transforming network cameras into powerful,data-generating sensors.By integrating AI-enabled edge devices,such as cameras,with large-scale computational power like cloud computing,the video surveillance industry is being redefined.Experts highlight this paradigm shift in the inter

44、views and report that AI-powered video solutions are increasingly being implemented in sectors like retail,logistics,and manufacturing.Common use cases include space optimization,people counting,asset tracking,and resource management.A recurring theme in these discussions is AIs ability to deliver s

45、ignificant cost savings and operational improvements.By integrating video analytics with broader business systems,companies can achieve better resource allocation,automate processes,and make more informed decisions.5KEY FINDINGS 9 insightsTHE STATE OF AI IN VIDEO SURVEILLANCE|12AI integration challe

46、nges:bridging expectations and capabilitiesAcross multiple interviews,a recurring theme emerged:the challenge of effectively integrating AI technologies to build systems that deliver better value.This is not just about leveraging technology for its own sake but about creating solutions that address

47、specific needs and generate tangible benefits.While AI offers significant potential to enhance surveillance and detection accuracy,several experts say they face difficulties in combining the right tools and technologies to produce reliable results.Selecting the optimal combination of AI software and

48、 hardware to meet customer needs efficiently remains a complex task.Advancements in AI hardware have paved the way for new capabilities,but they also introduce challenges,particularly for integrators working to ensure compatibility with existing legacy systems.6Just as we have a responsibility to ta

49、lk about the opportunities AI can bring,we also need to be realistic about what the technology can and cannot do.Partner,EMEASince almost anything is possible with AI,the real challenge is helping customers identify the right problems to solve where theyll get a real return on investment.Partner,Ame

50、ricasAnother critical issue is the gap between client expectations and the actual capabilities of AI solutions.Customers often anticipate near-perfect accuracy especially for applications like license plate recognition or anomaly detection.However,AI systems may typically achieve accuracy of 85-95%u

51、nder some conditions,which may not align with these expectations.Factors such as video quality and environmental variability,also play a role in determining performance and the fact that different use cases inherently have different accuracy.This challenge seems to be particularly evident in facial

52、recognition and motion detection systems,where transparency about the strengths and limitations of AI technologies is essential.Experts emphasized the importance of managing expectations and fostering trust by clearly communicating what AI can and cannot achieve.KEY FINDINGS 9 insightsTHE STATE OF A

53、I IN VIDEO SURVEILLANCE|13Partnership and collaboration for effective and responsible AI integrationThe ethical challenges associated with the rapid deployment of AI technologies underscore the importance of choosing the right partners.Several interviewees stressed that selecting collaborators with

54、aligned values especially in areas like privacy,compliance,and view on AI ethics is essential for them.Partnerships that emphasize alignment on values and goals also play a critical role in overcoming the hurdles of AI integration.Collaboration among AI developers,integrators,and end users was repea

55、tedly identified as a key factor for enhancing the reliability,relevance,and long-term success of AI implementations.By involving customers early in the process,stakeholders can ensure that the technology(and the intended use of it)is better understood,and user feedback is incorporated.One practical

56、 example,highlighted by respondents across regions,is the concept of collaborative surveillance projects.In these initiatives,customers actively participate in the design and ongoing feedback loops of AI systems.This hands-on approach results in tailored solutions that address specific needs while k

57、eeping ethical considerations and real-world applicability in focus.Navigating responsible AI Interviewees emphasized that responsible AI and ethical considerations are critical priorities in the development and deployment of AI systems.Many expressed concerns that decisions based on AI could become

58、 biased or unreliable.Other risks could be connected to privacy violations,where facial recognition and behavior recognition may create both legal and ethical challenges.A recurring theme among respondents was the importance of embedding responsible AI practices early in the development process.This

59、 means designing systems that prioritize fairness,transparency,and data privacy from the outset,rather than integrating these considerations after deployment.The most significant risk highlighted by many was the proliferation of biased AI,which could have harmful societal consequences if not properl

60、y managed especially when applied in sensitive applications like surveillance.Several interviewees pointed to the EU AI Act as a pivotal regulatory framework shaping the responsible use of AI,particularly in high-risk areas.Most respondents viewed the Act positively,welcoming the clarity and predict

61、ability it brings.Several interviewees also called for increased regulation in other parts of the world to ensure that responsible AI practices are mandated by law on a global scale.The ethical use of facial recognition was another area of discussion,particularly in regions with stringent privacy la

62、ws.Many customers in these areas remain hesitant to adopt facial recognition technologies.While regulation was broadly acknowledged as necessary to build trust and accountability,several interviewees also stressed the need for a balanced approach to safeguard innovation while addressing critical pri

63、vacy and data security concerns.7I think that ethics in partnerships shouldnt be taken lightly.Weve turned down customers who wanted to use our technology in ways we felt were unethical.Choosing the right partners and the right customers is also a part of acting responsibly.Partner,EMEA 8KEY FINDING

64、S 9 insightsTHE STATE OF AI IN VIDEO SURVEILLANCE|14The road ahead -future trends to watchThe rapid pace of technological development in AI makes it increasingly difficult to pinpoint specific future trends.As several interviewees pointed out,the speed of innovation often makes trend forecasting see

65、m impossible.Instead,there seems to be a collective focus on adaptability and continuous learning as technology evolves.Ultimately,while the future is difficult to predict,the direction seems clear:AI solutions are not only becoming more powerful,they are also increasingly efficient,and more integra

66、ted with a broader ecosystem of technologies.Generative AI continues to transform security operations and enhance decision-makingGenerative AI has emerged as a significant area of interest,and this was also mirrored in the survey results.While several interviewees highlighted the risk that expectati

67、ons of generative AIs capabilities are inflated in some areas,the technology was broadly seen as offering promising applications,especially in security and safety.For operators working in security,generative AI will provide tools that support situational awareness and improve their ability to respon

68、d effectively to incidents.Over time,it will make automation of complex processes possible,improving operations,enhancing decision-making,and aiding in the planning,scoping,deployment,and configuration of surveillance systems.9I have seen AI models achieve superhuman attention to detail.I would put

69、them head-to-head with a person,and AI would outperform them across thousands of cameras.Partner,AmericasMoving forward I think we will see a mix of both technological advancements combined with better practices.Partner,APACSmarter AI pipelines Deep learning and large neural network models continue

70、to be pivotal in advancing AI capabilities.However,its not just about scaling up the size of these networks:the focus seems to be shifting toward building smarter and more efficient AI pipelines.These pipelines aim to reduce computational power demands while maintaining or even improving accuracy.Th

71、ese efficiency gains can be critical for broadening AIs applicability,especially in resource-constrained environments or edge-based deployments.AI and IoT convergence Another trend is the integration of AI with IoT and sensor technologies,which is driving the evolution of surveillance systems.By com

72、bining video data with inputs from sound detection,motion detection,or other sensors such as temperature sensors,systems are becoming more intelligent and responsive.This convergence allows for more proactive,context-aware surveillance solutions capable of addressing complex security challenges in r

73、eal time.KEY FINDINGS 9 insightsTHE STATE OF AI IN VIDEO SURVEILLANCE|15Deep-dive:Responsible AI in video surveillance As AI continues to transform security,safety,business intelligence and operational efficiency,the conversation around responsible AI has gained increasing importance,raising fundame

74、ntal questions about privacy,ethics,and the potential for misuse.THE STATE OF AI IN VIDEO SURVEILLANCE|KEY FINDINGS Deep-diveTHE STATE OF AI IN VIDEO SURVEILLANCE|16 Facial recognition is highlighted as a very powerful AI,but also one of the most controversial.While it has the potential to enhance s

75、ecurity and safety,it can also lead to serious privacy violations if misused.Experts noted that in regions like Europe,where privacy laws are stricter,there seems to be a greater emphasis on ensuring that AI,including facial recognition technologies,are being used responsibly.Human oversight is key:

76、Several interviewees stressed the importance of maintaining human oversight in AI decision-making processes.AI can provide valuable assistance but should never replace human judgment to ensure that ethical considerations and context-specific nuances are taken into account.Given this,AI systems must

77、be designed to support,rather than replace,humans.Transparency a prerequisite for trust:Transparency is pointed out as a key to building trust for AI systems.Customers and end users need to understand how AI operates,what data it collects,and how decisions are made.Without transparency,there is a si

78、gnificant risk of distrust,especially when it comes to sensitive applications like facial recognition.Being transparent about systems limitations and capabilities is an important part of this.One of the central themes emerging from the expert interviews is the need for responsible AI practices.Some

79、of the key topic areas that were brought up:Responsible AI also means being transparent with customers about what the system can and cannot do.Its about setting the right expectations and ensuring that AI enhances security without compromising privacy.Partner,APACKEY FINDINGS Deep-diveTHE STATE OF A

80、I IN VIDEO SURVEILLANCE|17 Data privacy and security are major concerns in using AI for surveillance,especially in regions with strict data protection laws.Interviewees emphasized the importance of handling personal data,with great care.To meet privacy regulations and ethical standards,experts agree

81、d that companies must also focus on minimizing data collection and anonymizing data whenever possible.These practices help reduce the risk of privacy breaches while safeguarding sensitive information.Many interviewees also emphasized the importance of privacy by design,advocating for AI systems that

82、 limit the need for personal data.For instance,systems could be developed to analyze patterns or behaviors rather than identifying individuals directly,achieving effective results while minimizing the collection of identifiable information.With AI systems increasingly relying on large datasets and n

83、etworked infrastructure,data security is paramount.Systems must be designed to protect against cyberattacks and unauthorized access to sensitive data.This includes adequate encryption,secure storage practices,and strict access controls,just to mention a few of the measures needed.AI enables human-au

84、gmented workflows filtering out noise and allowing humans to focus on what actually matters.Partner,AmericasWe always say that AI should act as an aid,not make the final critical decisions.A human in the loop is essential,especially in security where the stakes are high.Partner,EMEAKEY FINDINGS Deep

85、-diveTHE STATE OF AI IN VIDEO SURVEILLANCE|18 Bias and fairness in AI systems:AI systems,particularly those involving facial recognition and predictive analytics,are vulnerable to bias if not developed and tested carefully.This is especially important for the safety and security applications where b

86、iased AI can lead to false positives or unfair outcomes for individuals,such as misidentification or unequal treatment of individuals.Interviewees emphasized the need for more diverse training datasets and rigorous testing to ensure that AI systems are fair and unbiased.Ongoing monitoring and testin

87、g:Responsible AI also requires continuous monitoring and testing to ensure that systems remain fair,accurate,and unbiased.This involves regularly updating algorithms with new data and conducting audits to identify and address any potential biases.Compliance and governance in focus:As AI technologies

88、 evolve,so do the regulations governing their use.Experts highlighted that a key factor for companies deploying AI in safety and security is to stay ahead of regulatory changes and be meticulous in ensuring that their systems comply with existing and emerging laws.We are excited about edge AI.Beside

89、s improving performance,privacy is enhanced when you are not passing data through multiple different systems.Partner,APACThe EU AI Act will shape how we deploy AI,particularly in high-risk areas like facial recognition.Companies will need to invest more in compliance to ensure their systems meet the

90、se new requirements and those that will follow.Partner,EMEA Building public trust in AI is an industry-wide responsibilityPublic trust is the cornerstone of any industrys license to operate.In-depth interviews highlight that widespread adoption of AI depends on maintaining the confidence of the publ

91、ic,authorities,and policymakers.This trust must be built through responsible AI practices adopted across industries,not just by individual companies.KEY FINDINGS Deep-diveTHE STATE OF AI IN VIDEO SURVEILLANCE|19ConclusionsThe use of AI in security,safety,business intelligence,and operational efficie

92、ncy is growing rapidly,with both partners and end customers recognizing its potential and the opportunities it brings.Despite the advancements made in recent years,there remains caution around the technologys limitations and risks.A clear understanding of what AI can realistically achieve,how new re

93、gulations set the scene for future innovation,and what responsible AI means in practice is needed.THE STATE OF AI IN VIDEO SURVEILLANCE|ConclusionsTHE STATE OF AI IN VIDEO SURVEILLANCE|20The key pillars of this approach are:Human-centric with respect for human rights:AI technology should serve to en

94、hance human capabilities and contribute to a better world.AI must be designed to ensure safety and respect for people.Responsibility and accountability:We are committed to building the world we want to live in,and we acknowledge the societal impact of our work.This means that we develop,deploy,and u

95、se AI responsibly in our operations and across our value chain.We are accountable for the AI systems we create and to ensure they follow ethical principles,comply with laws,and manage risks.Responsible innovation:We never inhibit explorative technology development because we believe that overlooking

96、 opportunities carries the risk of missing out on enhancements that could benefit individuals and society.However,there is a significant difference between exploring new ideas and commercializing them.We only commercialize innovations that we ethically stand behind and are in line with our vision of

97、 innovating for a smarter,safer world.A risk-based approach:We understand that different AI technologies carry different risks.To fully grasp these risks,potential applications and use cases must always be considered before new products and solutions are made available to others.All our products and

98、 solutions undergo thorough assessment before being made available to the market.Embedded ethics:Technology will invariably outpace legislation.Given this,emphasizing ethics within culture and practices is increasingly crucial for us as responsible innovators.While ethics are deeply embedded into Ax

99、is core values,culture and ways of working,genuine responsibility also requires clear processes,governance,and transparency.Knowledge-led leadership based on partnership and collaboration:As industry leaders,we believe it is our duty to shape safe and ethical AI practices within our industry and to

100、facilitate an informed discourse on AIs capabilities,strengths and limitations and ethical dilemmas.This responsibility involves educating and sharing our expertise and learnings with others from partners and end customers to political decision-makers,civil society,and media.Axis approach to AIAxis

101、approach to AI is rooted in one overarching principle:That AI technology,just like all technologies,should leverage and augment human intelligence,build on respect for human rights and should benefit people and society.This is in line with Axis vision,to innovate for a smarter,safer world.As a compa

102、ny,we are committed to the advancement of AI driven by ethical principles.We stand by the OECD Principles on Artificial Intelligence,prioritizing transparency,trustworthiness,and the protection of human rights and democratic values.AI remains one of the most powerful and transformative technologies

103、within the video surveillance industry.While there are significant opportunities for AI to improve safety and security,operational efficiency and business intelligence,there must be a focus on ethical implementation and meaningful integrations which drive value.Mats Thulin|Director of AI and Analyti

104、cs Solutions,Axis CommunicationsAxis approach to AITHE STATE OF AI IN VIDEO SURVEILLANCE|21APPENDIXAppendix 1:About Axis end customer survey and Axis sales channel surveyThe annual Axis sales channel survey and the Axis end customer survey are two global studies conducted by Axis during Q2 2024.The

105、Axis sales channel survey targeted distributors and channel partners across 68 countries,resulting in 4,900 responses.Meanwhile,the Axis end customer survey was sent to end customers in 64 countries,gathering approximately 900 responses.Both surveys included questions designed to explore impactful t

106、rends expected to influence respondents in the near future.For end customers,the question read:“Which of the following trends do you consider most significant for your business in the near future?”Respondents were allowed to select up to three options.For distributors and channel partners,the questi

107、on read:“Which of the following trends do you think will be most important for our industry in the near future?”Respondents were allowed to select multiple answers.The available answer options were identical for both respondent groups:AI/Generative AI Environmental sustainability Ethics and trust Cy

108、bersecurity,risk,and privacy Cloud solutions Analytics and actionable insights IoT and interconnectedness of physical devices Digitalization(moving existing processes into digital technologies)Supply chain efficiency No opinionAppendixTHE STATE OF AI IN VIDEO SURVEILLANCE|22Appendix 2:About the in-d

109、epth expert interviewsThe in-depth interviews for this report were conducted with selected experts from Axis partner network during 2024.The purpose of these interviews was to gain qualitative insights into the current landscape of experts working with AI to improve security,safety,business intellig

110、ence or operational efficiency as well as to explore the opportunities,challenges,and ethical considerations associated with AI integration in this area.The interview process involved the following steps:1.Participant selection:Experts were carefully selected based on their knowledge and experience

111、in AI to ensure that a diverse range of perspectives was captured.2.Interview format:Each interview was semi-structured,allowing for flexibility in discussion while ensuring that key topics were covered.This format encouraged participants to share their insights and experiences in depth.3.Data colle

112、ction:Interviews were conducted through online meetings,in person sessions,or phone calls,depending on participant availability.Most sessions were recorded and then transcribed for analysis.4.Analysis:The transcriptions were analyzed to identify key themes,trends,and insights,which were then integra

113、ted into the report.The interview guide included the following questions:Can you briefly introduce yourself and your role within your organization?For how long have you been working in the security and surveillance industry?What is your experience with AI technologies?How are you working with AI tod

114、ay?What do you perceive as the primary opportunities for utilizing AI in video surveillance to enhance safety,security,and beyond?Which technologies do you see most potential in?-Gen AI,LLM,foundational models,Video Analytics,Machine Learning,NLP,Natural Language Processing,Biometrics,Predictive Ana

115、lytics -Semantic learning and symbolic learning-Ranked according to potential,long-term vs short term What are the key challenges associated with the integration of AI in these areas?What are the biggest risks associated with the integration of AI in these areas?What do you do to mitigate these risk

116、s?How do you envision AI evolving in the safety,security,and surveillance industry in the next year?And in 5-10 years?How do you perceive your market in relation to this development?What does responsible AI mean to you in the context of the security and surveillance industry?Do you see any risks in

117、regard to ethics?Do you know any good examples of responsible AI initiatives from the industry?Are these from within your organization?What do you believe will be the most significant developments or trends in AI applications in the coming years?How do you think the industry can address the challeng

118、es tied to AI discussed earlier?How do you see Axis role in this?Appendix2025 Axis Communications AB.AXIS COMMUNICATIONS,AXIS,ARTPEC and VAPIX are registered trademarks of Axis AB in various jurisdictions.All other trademarks are the property of their respective owners.Axis enables a smarter and saf

119、er world by improving security,safety,operational efficiency,and business intelligence.As a network technology company and industry leader,Axis offers video surveillance,access control,intercoms,and audio solutions.These are enhanced by intelligent analytics applications and supported by high-quality training.Axis has around 5,000 dedicated employees in over 50 countries and collaborates with technology and system integration partners worldwide to deliver customer solutions.Axis was founded in 1984,and the headquarters are in Lund,Sweden.About Axis Communications