Opinion by: Merav Ozair, PhD
The launch of ChatGPT in late 2023 sparked an arms race amongst Huge Tech firms equivalent to Meta, Google, Apple and Microsoft and startups like OpenAI, Anthropic, Mistral and DeepSeek. All are dashing to deploy their fashions and merchandise as quick as doable, asserting the subsequent “shiny” toy on the town and attempting to say superiority on the expense of our security, privateness or autonomy.
After OpenAI’s ChatGPT spurred main progress in generative AI with the Studio Ghibli pattern, Mark Zuckerberg, Meta’s CEO, urged his groups to make AI companions extra “humanlike” and entertaining — even when it meant enjoyable safeguards. “I missed out on Snapchat and TikTok, I received’t miss out on this,” Zuckerberg reportedly stated throughout an inner assembly.
Within the newest Meta AI bots venture, launched on all their platforms, Meta loosened its guardrails to make the bots extra participating, permitting them to take part in romantic role-play and “fantasy intercourse,” even with underage customers. Workers warned concerning the dangers this posed, particularly for minors.
They may cease at nothing. Not even the security of our kids, and all for the sake of revenue and beating the competitors.
The injury and destruction that AI can inflict upon humanity runs deeper than that.
Dehumanizing and lack of autonomy
The accelerated transformation of AI doubtless results in full dehumanization, leaving us disempowered, simply manipulable and fully depending on firms that present AI providers.
The newest AI advances have accelerated the method of dehumanization. We now have been experiencing it for greater than 25 years for the reason that first main AI-powered advice techniques emerged, launched by firms like Amazon, Netflix and YouTube.
Corporations current AI-powered options as important personalization instruments, suggesting that customers could be misplaced in a sea of irrelevant content material or merchandise with out them. Permitting firms to dictate what folks purchase, watch and suppose has change into globally normalized, with little to no regulatory or coverage efforts to curb it. The results, nonetheless, could possibly be important.
Generative AI and dehumanization
Generative AI has taken this dehumanization to the subsequent degree. It turned frequent apply to combine GenAI options into current purposes, aiming to extend human productiveness or improve the human-made end result. Behind this large push is the concept people usually are not ok and that AI help is preferable.
Current: Meta opens Llama AI mannequin as much as US navy
A 2024 paper, “Generative AI Can Hurt Studying,” discovered that “entry to GPT-4 considerably improves efficiency (48% enchancment for GPT Base and 127% for GPT Tutor). We additionally discover that when entry is subsequently taken away, college students carry out worse than those that by no means had entry (17% discount for GPT Base). That’s, entry to GPT-4 can hurt academic outcomes.”
That is alarming. GenAI disempowers folks and makes them depending on it. Individuals might not solely lose the flexibility to supply the identical outcomes but additionally fail to speculate effort and time in studying important expertise.
We’re shedding our autonomy to suppose, assess and create, leading to full dehumanization. Elon Musk’s assertion that “AI shall be manner smarter than people” is no surprise as dehumanization progresses, as we are going to now not be what really makes us human.
AI-powered autonomous weapons
For many years, navy forces have used autonomous weapons, together with mines, torpedoes and heat-guided missiles that function based mostly on easy reactive suggestions with out human management.
Now, it enters the world of weapon design.
AI-powered weapons involving drones and robots are actively being developed and deployed. Resulting from how simply such expertise proliferates, they’ll solely change into extra succesful, subtle and extensively used over time.
A significant deterrent that retains nations from beginning wars is troopers dying — a human value to their residents that may create home penalties for leaders. The present growth of AI-powered weapons goals to take away human troopers from hurt’s manner. If few troopers die in offensive warfare, nonetheless, it weakens the affiliation between acts of struggle and human value, and it turns into politically simpler to start out wars, which, in flip, might result in extra loss of life and destruction total.
Main geopolitical issues may shortly emerge as AI-powered arms races amp up and such expertise continues to proliferate.
Robotic “troopers” are software program that is perhaps compromised. If hacked, all the military of robots might act in opposition to a nation and result in mass destruction. Stellar cybersecurity could be much more prudent than an autonomous military.
Keep in mind that this cyberattack can happen on any autonomous system. You may destroy a nation just by hacking its monetary techniques and depleting all its financial sources. No people are harmed, however they could not be capable of survive with out monetary sources.
The Armageddon state of affairs
“AI is extra harmful than, say, mismanaged plane design or manufacturing upkeep or unhealthy automobile manufacturing,” Musk stated in a Fox Information interview. “Within the sense that it has the potential — nonetheless small one might regard that likelihood, however it’s non-trivial — it has the potential of civilization destruction,” Musk added.
Musk and Geoffrey Hinton have not too long ago expressed issues that the potential of AI posing an existential risk is 10%-20%.
As these techniques get extra subtle, they could begin performing in opposition to people. A paper printed by Anthropic researchers in December 2024 discovered that AI can pretend alignment. If this might occur with the present AI fashions, think about what it may do when these fashions change into extra highly effective.
Can humanity be saved?
There’s an excessive amount of give attention to revenue and energy and virtually none on security.
Leaders needs to be involved extra about public security and the way forward for humanity than gaining supremacy in AI. “Accountable AI” isn’t just a buzzword, empty insurance policies and guarantees. It needs to be on the high of the thoughts of any developer, firm or chief and applied by design in any AI system.
Collaboration between firms and nations is vital if we want to forestall any doomsday state of affairs. And if leaders usually are not stepping as much as the plate, the general public ought to demand it.
Our future as humanity as we all know it’s at stake. Both we guarantee AI advantages us at scale or let it destroy us.
Opinion by: Merav Ozair, PhD.
This text is for common info functions and isn’t meant to be and shouldn’t be taken as authorized or funding recommendation. The views, ideas, and opinions expressed listed here are the creator’s alone and don’t essentially replicate or signify the views and opinions of Cointelegraph.