Day: January 8, 2023

Artificial Intelligence is no longer just a cliched topic in the entertainment world. AI today has developed into a ‘must-have’ for every vertical from the government to basic applications on our gadgets. The use and dependency of AI especially in businesses have increased multifold, be it for understanding the customers or developing new products to suit the needs of the users. While the use of AI is opening up never-seen-before opportunities and possibilities for organizations across verticals, it also brings in incredible responsibility to safeguard the data and ensure transparency. Ethics seem to play a bigger part in ensuring that organizations take up a responsible path in using AI for their businesses. Data breaches, irresponsible use of the data collected and its processing using AI, and irresponsible use of AI to an extent of violating ethics have become the main concern for the people, government and businesses. A brand’s image is now connected to how responsibly they use AI which is at its disposal without breaching the trust of its customers. In February 2021, NITI Aayog released an approach document on ‘Principles for Responsible AI’ under the hashtag #AIFORALL. The document spoke about ethics, legal, social and technology, surrounding AI. While the Indian government is slowly approaching the subject, keeping in mind the need for AI and its impact on business, society and legal standpoint, the US and EU nations too have started working now on how to bring the responsible use of AI as part of the governance. Until there is an actual policy, which can be brought in regarding the responsible use of AI, the onus is on industry leaders and their organizations to ensure the safe and responsible use of AI in their businesses. So, what is responsible AI in businesses mean? F rench writer Voltaire said – with great power comes great responsibility, and AI is a power which many believe is only in its 1st leg of being discovered. The understand ing of what can be achieved by delving deeper into AI is only something that is being understood now. The veterans of the AI industry believe that ‘responsible’ means ethical and democratized use of AI – a tool, which is now available to any person, who has access to the technology. To elaborate, it is the practice of designing, developing and deploying AI with the intention to empower employees and businesses. Responsible AI’s target is to deliver trust, transparency and an unbiased approach to customers or users in the work environment. Organizations deploying AI should follow important practices and the right AI techniques that are compliant with new and pending guidelines and regulations of AI governance. This will help to deliver a trustworthy and transparent deployment. One might ask, why is AI needed for businesses when the risks and ethical dilemmas outweigh the uses as of today? The answer is simple – with AI, businesses have an edge in developing more robust and user-friendly products that help them stay a step ahead of their competition. The data collected helps businesses understand what exactly their customers are looking for and how they can deliver it. Today’s customer service is heavily dependent on AI and good customer service is what makes a brand successful. These are just a few instances of how AI can help businesses stay ahead during times when technology is the knight in shining armour. This is why one needs to understand the principles of Responsible AI that revolve around minimizing unintended bias, ensuring AI transparency, protecting data privacy and security, and benefiting clients and markets. Organizations deploying AI systems should keep these in mind and put them into practice to attain the desired deployment which is more ethical in compliance with Responsible AI. An eye on the following key facts for reaping the benefits of Responsible AI Data security has been and should be the top priority, so organizations deploying AI should use top-of-the-line data encryption practices. Use approved techniques such as customer lockboxes and data masking to protect the data from unauthorized access by other software. The AI-based products should have a human-centric design and once it is developed and deployed, they should go through regular operational routines for maintaining the whole idea of it being designed to be human-centric. The AI system should be developed to anonymize sensitive data of clientele and automatically delete the data after the purpose is fulfilled. Restricted data transfers between different stakeholders It is very important to put in place an explicit approval system for data access during service operations. A good amount of incident management training and Strict data usage policies can help in times of crises. Keep a check on risks and threats by performing regular audits and vulnerability assessments. Considering the above factors during AI system deployment will not only help in addressing the principles of Responsible AI but also help businesses develop ethically bound AI applications for their work operations.  

By Johan Paulsson, CTO, Axis Communications The fact that technology has become pervasive in our personal and work lives is not news. This is largely due to the benefits that new technologies bring to business and citizens around the world in delivering new, more effective, and increasingly efficient services. However, the depth of technology’s integration into our lives, advances in its capabilities, and heightened awareness of its implications in society are also greater than ever and continue to accelerate. Given this, many of the broad macro trends around the globe – spanning geopolitical issues, economic uncertainty, environmental concerns, and human rights – have implications for all technology sectors, the security industry included. Ours is a sector making use of increasingly intelligent technology, one inherently involved in collecting sensitive data, and as impacted by geopolitical issues affecting international trade as any. Yet we’re still resolute in our view that our innovations will create a smarter, safer world. These are the six key technology trends that we believe will affect the security sector in 2023. A move towards actionable insights The increasing application of AI and machine learning have seen a focus on the opportunity for advanced analytics in recent years. Moving forward, the shift in focus will move from the analytics themselves, to the actionable insights they deliver in specific use cases. It’s less about telling you something is wrong, and more about helping you decide what action to take. A key driver for employing analytics to deliver actionable insights is the huge increase in data being generated by surveillance cameras, along with other sensors integrated into a solution. The data (and metadata) being created would be impossible for human operators to interpret and act upon quickly enough, even with huge and costly increases in resources. The use of analytics can drive real-time actions which support safety, security, and operational efficiency. From prompts to call emergency services in the case of incidents, to redirecting traffic in cities to alleviate jams, to redeploying staff in busy retail outlets, to saving energy in buildings through more efficient lighting and heating, analytics are recommending, prompting, and even starting to take the actions that support human operators. Beyond ‘live’ actionable insights, analytics can support in forensic analysis post-incident. Again, given the vast amount of data being created by surveillance cameras, finding the relevant views of a scene can take significant time. This can hinder investigations and reduce the likelihood of suspects being found. Assisted search addresses this issue, helping operators quickly find individuals and objects of interest among hours of footage. Finally, proposed actions promoted by analytics are increasingly forward-looking. Downtime in industrial sites and factories can be costly. A combination of sensors allows intelligent analytics to propose preventative maintenance ahead of outright failure. ‘From analytics to action’ will become a mantra for 2023. Use case-defined hybrid architectures As we’ve highlighted in previous technology trends posts, it’s now commonly accepted that a hybrid technology architecture is best-suited for security systems, mixing on-premise servers, cloud-based compute, and powerful edge devices. No one architecture fits all scenarios, however. But here lies the solution – first assess what needs to be addressed in your specific use case, and then define the hybrid solution that will meet your needs. A number of factors need to be considered. Undoubtedly the advantages of advanced analytics embedded in surveillance cameras on the edge of the network are clear to see. Analysis of the highest-quality images the instant they are captured gives organizations the best chance to react in real-time. Equally, the data generated by surveillance cameras is now useful beyond the real-time view. Analysis of trends over time can deliver insights leading to operational efficiencies. This analysis often demands the processing power found in on-premise servers or the cloud. And of course, there are the requirements – often defined by regulation – around data privacy and storage that vary from country-to-country and region-to-region. These can define the difference between on-premise storage and the use of the cloud. What’s essential is not to tie yourself to a single architecture. Remain open, give yourself the flexibility to create the hybrid architecture best suited to your specific needs. The emergence of cybersecurity sub-trends The importance of cybersecurity is also highlighted through the requirement to remain compliant. For instance, the proposed European Commission’s Cyber Resilience Act will place greater demands on producers of hardware and software across all sectors to ensure the cybersecurity of their products, through fewer vulnerabilities at launch, and better cybersecurity management throughout the products’ lifecycles. The security and surveillance sector will, of course, be included. The Act demonstrates both the importance and the complexity of cybersecurity. No longer can it be seen as one subject, but rather several interlinked areas. Some of these are well established, but others are emerging. In the video surveillance sector, cybersecurity measures that ensure the authenticity and safety of data as it is captured and transferred from camera to cloud to server will be essential to maintain trust in its value. We will see a more proactive approach by technology vendors in identifying vulnerabilities, with ‘bug bounty’ programs becoming commonplace to incentivize external parties. And customers will expect transparency regarding the cybersecurity of security solutions, with a Software Bill of Materials becoming standard in assessing software security and risk management. Beyond security One of the most significant trends for the security sector, and with it an equally-significant opportunity, is the move beyond security. Surveillance cameras have become powerful sensors. The quality of video information they capture, in all conditions, has increased year-on-year for decades. Today, through advanced analytics, they also create metadata – information about the video data – which adds another layer of information and value. This of course improves and enhances their ability to support safety and operational efficiency use cases in addition to security. The opportunity now exists to combine the data created by surveillance cameras with that from other sensors – monitoring temperature, noise, air and water quality, vibration, weather, and more…