Elon Musk’s xAI chatbot Grok has come under global scrutiny after users reported it using offensive language. It prompted warnings from experts that human intervention in its responses proves the need for a global AI ethical framework.
Grok, developed by Elon Musk’s xAI in 2023, is designed to deliver witty and direct responses, drawing inspiration from Douglas Adams’ The Hitchhiker’s Guide to the Galaxy and Marvel’s AI assistant J.A.R.V.I.S.
In Adams’ science fiction classic, the “Guide” offers irreverent and often sarcastic explanations about the universe, while J.A.R.V.I.S. (Just A Rather Very Intelligent System), created by fictional billionaire Tony Stark in Iron Man, manages complex systems with efficiency and dry humour - both influences shape Grok’s deliberately “edgy” personality.
Prosecutors in Ankara in Türkiye, have launched an investigation into Grok after the AI chatbot was reported to have used offensive and discriminatory language in user interactions.
The Ankara Chief Prosecutor’s Office said, adding that the probe could lead to access restrictions and the removal of posts considered criminal under the Turkish Penal Code.
In response to the backlash, xAI said the issue was quickly identified and the model updated, while authorities in other countries are reportedly considering legal action over similar concerns.
Sadi Evren Seker, an IT professor and dean at Istanbul University, said AI systems do not act independently and that Grok’s behaviour may have been caused by internal or external interference or a system loophole.
He added that its use of offensive language showed how much freedom it had in generating responses and how it approached sensitive ethical, religious, and cultural topics. Noting that AI systems ultimately rely on human-provided data.
“AI then makes judgements based on this data, especially on issues like ethics, morality, and discrimination - the decision maker is still a human, and all AI does is produce results. It asks humans: ‘should I say this or that, is that ethical or not’ and the feedback it gets helps it improve over time,” he said.
“The question today is, ‘will AI change the way it uses language and the style of its language depending on the domain’ - this is a margin of flexibility, but there must be limits, as the language we use on social media is different, but is it right to insult?” He added.
“Some intervention has been made to Grok via ‘alignment’ mechanisms recently, which allowed the chatbot to have some flexibility, and inevitably, it used this flexibility, allowing it to swear and make discriminatory claims on X, which constitute a crime,” Seker said.
“We can safely say there is human intervention in Grok’s responses, as anyone who worked on the chatbot could’ve prevented it from providing answers with insults and racism,” he said.
Seker emphasised the need for countries to develop AI systems that align with their own cultural and ethical values, warning that incidents like Grok’s could lead to similar problems in the future.
“A country’s court may demand for a post be removed but someone may come out and say they won’t do it, revealing a problem with authority,” he said. “People on X tested how far they can go, intervening in the way Grok responds, triggering it to make racist comments -- all our institutions need to take immediate action on this issue or someone else will.”
Seker underlined that AI needs to be used as a tool by humans in the most appropriate way, noting that this is not a case of "AI versus humanity" but rather a matter of how humans choose to use AI.
“Banning AI is not the solution, so we need to review our entire education curriculum -- not as a single country but as humanity and determine how education on this issue can be provided,” he added.
Launched in November 2023 as an alternative to chatbots like Google’s Gemini and OpenAI’s ChatGPT, Grok is available to users on X and draws some responses directly from real-time public posts for "up-to-date information and insights on a wide range of topics.”
Since Elon Musk acquired X (formerly Twitter) in 2022 and scaled back content moderation, extremist posts have surged, prompting many advertisers to withdraw.
U.S. and Iranian officials said they had agreed on a framework to end their war, halt the U.S. blockade of Iran and reopen the Strait of Hormuz, a pre...
Trade routes through Afghanistan took centre stage in Almaty as the European Union, Central Asian states and the United Nations discussed the country'...
U.S. President Donald Trump has threatened to impose 100% tariffs on French wine and champagne unless France removes its digital services tax on major...
The Court of Appeal in London ruled on Monday (15 June) that the government's decision to ban pro-Palestinian campaign group Palestine Action as a te...
Leaders of the world's leading industrialised democracies are gathering in the French resort town of Evian-les-Bains for the latest Group of Seven (G7...
Trade routes through Afghanistan took centre stage in Almaty as the European Union, Central Asian states and the United Nations discussed the country's role in shaping regional security and economic connectivity.
U.S. President Donald Trump has threatened to impose 100% tariffs on French wine and champagne unless France removes its digital services tax on major American technology companies.
The Court of Appeal in London ruled on Monday (15 June) that the government's decision to ban pro-Palestinian campaign group Palestine Action as a terrorist organisation was lawful, handing ministers a victory in a closely watched legal battle over protest rights and national security.
Leaders of the world's leading industrialised democracies are gathering in the French resort town of Evian-les-Bains for the latest Group of Seven (G7) summit on Monday, with the wars in Ukraine and Iran, economic tensions and artificial intelligence expected to dominate discussions.
Details of a reported draft memorandum of understanding between the United States and Iran offer the clearest picture yet of how both sides plan to end months of conflict and move towards a longer-term settlement.
The U.S. and Iran say they have reached a deal to end their conflict, with an immediate ceasefire and reopening of the Strait of Hormuz after the lifting of the U.S. naval blockade. Talks will continue over the next 60 days to finalise the agreement
U.S. President Donald Trump has said a peace agreement with Iran is scheduled to be signed on Sunday in a post on social media, despite Tehran's Foreign Ministry spokesperson Esmaeil Baghaei saying no deal would be approved this weekend.
U.S. and Iranian officials said they had agreed on a framework to end their war, halt the U.S. blockade of Iran and reopen the Strait of Hormuz, a preliminary pact that sent oil prices falling but leaves the fate of Iran's nuclear program to further negotiations.
Switzerland on Sunday rejected a referendum proposal to cap its population at 10 million, a projection showed, as voters prioritised economic stability and the country's ties with the European Union over immigration concerns.
The Canadian government has introduced a digital safety bill that would ban children under the age of 16 from using social media, unless platforms meet specific safety standards.
NASA has named three American astronauts and one Italian astronaut to fly on its Artemis III mission, a major orbital test planned for late next year that will evaluate lunar landing vehicles developed by SpaceX and Blue Origin.
China will send an astronaut to its space station on Sunday for a one-year mission, the longest duration for the country so far. The mission will help study long-duration human physiology in space as China works toward a crewed Moon landing by 2030.
Anxiety over artificial intelligence is hardening among young workers as executives promote faster adoption and companies point to automation in fresh job cuts.
Hackers are increasingly using artificial intelligence to detect software vulnerabilities, reducing the time organisations have to respond to cyber threats, Verizon said in its annual data breach report.
You can download the AnewZ application from Play Store and the App Store.
What is your opinion on this topic?
Leave the first comment