The challenge facing governments internationally as they attempt to hold social media and artificial intelligence companies to account was evident on several fronts this week.
The fundamental antagonism these businesses have to the idea of official oversight of their activities was there for all to see at the Oireachtas media committee, which quizzed executives from TikTok, Meta and Google about removing content promoting illegal activity. It followed the proliferation of dangerous driving videos which came into focus in the aftermath of the M9 crash last month.
The yawning gap between the professed policies of these companies on removing such content and their enthusiasm for doing so reflects their deep ambivalence to the idea of any sort of regulation. One platform, X, did not even make an appearance.
The committee held out the prospect of social media company directors being held criminally liable for online safety failures given the apparent ineffectiveness of levelling fines.
READ MORE
The idea is unlikely to gain much traction, but the weakness of the current fines based approach was also illustrated this week when the Data Protection Commissioner fined Google €403 million for breaching data privacy rules about the processing of users’ location history and data.
Substantial though it may be, the fine is far from material for Alphabet – Google’s parent. More pertinently it relates to events which occurred back in 2018 and the issue is now moot.
It will most likely join the approximately €3.5 billion in fines that have been levied against large social media companies by Irish regulators, the bulk of which has yet to be paid because of ongoing legal wrangling.
The inadequacy of the tools to rein in social media giants and the libertarian culture of the companies has resulted in a regime that always appears to be– and often is – at least one step behind its charges.
This does not bode well for the significant danger posed by artificial intelligence. The risks posed by weak or non-existent regulation are compounded by the increasing possibility that AI companies are losing control of their creations. The revelation on Thursday that an autonomous AI agent built by OpenAI hacked the Australian public health service without the knowledge of its creator is the latest warning of the growing risks presented by “runaway AI”.
The need for some sort of global pause in development as suggested by some of the main players – but opposed by the US government – is clear. An international regulatory structure is needed, though looks unlikely to emerge. What is also clear is the need for a fundamental shift in the mindset of these businesses regarding accountability to the societies in which they prosper.
