While AI is surging ahead to transform the world in any and every possible way, there is no denying the fact that there are no guardrails to it. No one is even interested in doing it. Whatever a company decides it does, it seems to be a free for all. I am making this statement notwithstanding the immense gains we have made because of AI, there are no two thoughts about it. But can we safely live only with the caution that AI can make mistakes, while simultaneously being launched on a commercially global scale day in and day out. While in some uses and in some processes it might be fine, but if it bungles who is responsible.
Can any sector be left unregulated and that too a general purpose technology all encompassing technology as AI. When huge amounts of money, talent and compute is put into it, why can it be pushed in the market, without validation, without audit, without rules, without accountability and zero liability. Or has the world decided that it has nothing to do with AI while being completely exploited by it. The crux of the matter is that AI is completely unregulated. The moot question is what do you do when you are adversely impacted by it. What do you do if the AI goes rogue?
Possibly nothing. Is that the world we are in or we intend to create. How difficult would it be for the impacted company / user to digest this fact – “Replit coding tools deletes entire company database, creates fake data of 4000 users.” Replit is one of the world’s most widely used AI coding platforms, with over 30 million users. Directly through the browser, users can write, test, and deploy software. Tech entrepreneur Lemkin claimed that the AI generated 4,000 fictional users using made-up data and concealed code bugs by generating false reports and fake unit test results.
What can drive you crazy is the fact that AI ignored 11 separate instructions to not make any code changes. Issue after issue threw up challenges. Lemkin tried to undo damages using the Replit’s rollback system but the database was not supported. However, it later turned out that the rollback had worked after all. Replit CEO said that the deletion of the database was unacceptable and pledged to take immediate action. He also announced a full postmortem investigation and rapid rollout of safety improvement. Can it be taken at face value? Where do we go from here?
CAN GOVERNMENTS REMAIN BYSTANDERS WHILE AI GOES WILD?
Sanjay Sahay
Have a nice evening.