Understanding AI’s Ministry of Truth

AI has brought the technological landscape to a crossroads. It promises immense productivity gains and revolutionized work, yet poses challenges tied to manipulation, data privacy and security. From deepfakes to falsified social media posts, leaders across the globe are realizing the true scope of AI’s consequences if not mandated properly. Critical to this juncture are regulations. However, governments hesitate to introduce AI regulations for fear of stifling innovation or enacting counterintuitive, overly complex rules.

Initial small but meaningful steps toward reining in AI have emerged, such as the Biden Administration’s industry commitments and the EU AI Act. The real shift to reckoning with AI’s power came in late September of this year when California Governor Gavin Newsom signed multiple bills to address sexually explicit deepfakes, AI watermarking and deepfake election content. These bills marked the first wide-ranging mandatory AI legislation adopted at the state level. A bellwether for the nation and world, reverberations are sure to grow and illustrate the importance of AI regulation. Governments will likely implement more restrictive cyber resilience, compliance and AI standards across the board, driven by the rapidly increasing number of cyberattacks and regulatory gaps.

As AI developments and regulation gain speed, a spotlight is simultaneously shone on data. Data’s value has skyrocketed with the advent of AI. Seen as the engine behind AI models, reams of reliable data are key to training and making the technology work. As a result, organizations must now consider the resiliency of their data to ensure that they can manage AI risks and opportunities while staying compliant. Here, two key pillars of data resilience will come into play: Data freedom and data backup.

With disparate regulations, data freedom becomes more challenging. Data freedom lets you move and store data wherever it’s needed – and is critical for businesses. You need to find a way to let everyone in who needs to get in, while keeping security in mind. Being a Broncos fan, it makes me think of NFL stadiums. Part of the magic trick their security pulls off is letting 80,000 into the stadium while keeping the event safe no matter where in the venue they move about – their seats, concessions, etc. The same is true for data freedom. Additionally, data freedom supports AI initiatives by ensuring that the right data is available, in the right location, as sources for artificial intelligence and machine learning activities – many of which utilize cloud-based services or require access to a wide range of data sets.

Giving access to corporate data but protecting your company’s most valuable assets is key. Much of security these days is about assuming “reasonable risk.” That is, what can we reasonably accept while maintaining safety standards – and AI is set to change this. Regulations suggest that what was a reasonable effort in the past may not be a reasonable effort in the future. Trying to secure data to ensure only authorized users can access it will place more pressure on security teams.

AI is a double-edged sword because it can be wielded by good and bad actors. With it being so easy to create deepfakes or falsify information, it’s worrisome for businesses, because AI has made it harder to discern truth from fiction. That’s why data resiliency – in the form of immutable backups – is now the only ministry of truth for AI. Following a ransomware attack or other cyber incident, authorities examine the network and data, treating it as a crime scene. Without data backup, there would be no way to prove the business took appropriate preventative measures to recover from attacks. After all, it’s not a case of if they’ll be impacted, but when.

Taking it a step further, imagine a job interview goes poorly. Looking for someone to blame, the candidate alleges the company is engaging in biased hiring practices by using an AI algorithm to assist in its search for candidates. The company knows it isn’t true, but in the age of AI, knowledge and belief are not enough. As this new frontier of AI continues to proliferate and improve, it will be the onus of businesses and individuals to prove their word against that of AI. Immutable, timestamped data backups are the only way to validate truthful information, and hence safeguard an organization’s integrity.

AI has changed how business operates. In giving the world a powerful productivity tool, it’s simultaneously created a difficult problem – how to prove something is true when anything can be fabricated at a moment’s notice. How can authorized users gain access to data while keeping unwanted actors out – all while ensuring the technology complies with the myriad regulations from states and countries? The key is finding a singular source, or ministry, of truth.

For businesses, that source is data resiliency.