When reports spread that X users were asking the platform’s artificial intelligence (AI) chatbot Grok to turn photographs of celebrities and nonpublic figures — including minors — into sexualized images, public outrage rightly focused on the harm inflicted on the victims. The majority of the content targets women, and it causes reputational damage and psychological distress.
There is another kind of damage that is being overlooked in the discourse: what this technology does to the people who create the images. This is not an attempt to summon sympathy for bad actors. It is worthy of attention because naming the self-inflicted costs could act as a much-needed deterrent.
Over the years, concerns about pornography’s ubiquity have been about how easy access and exposure might be negatively influencing sexual behavior and even eroding relationship bonds. What is happening with Grok and other tools should heighten those worries.
For all its raw immediacy, traditional pornography is still at arm’s length — a sexual fantasy typically acted out by consenting adults who are strangers to the spectator. AI-generated pornographic deepfakes can drastically narrow that distance.
Suddenly the viewer is the producer, and the images can turn a coworker, a barista or a date into an explicit simulation, blurring the line between fantasy and a real sexual partner.
In the process, what has been taught should be a respectful, reciprocal pursuit is replaced with a private shortcut that requires no consent.
Throughout history (or at least since women stopped being chattel), the fact that humans are biologically wired to want sex has helped drive the onerous emotional work of connecting to other people. That includes learning to communicate, tolerating the uncertainty and fear that comes with vulnerability and negotiating needs with another person. These are skills that require effort and mastery, but the prospect of a sexual and romantic connection has often been a powerful motivator.
When someone can generate an AI image of the person they want, looking exactly how they want them to look and doing exactly what they want them to do without their consent, it encourages the technology’s user to bypass those building blocks. They have essentially gotten the “reward,” while skipping the work that is essential for forming lasting relationships offline, training themselves, click by click, to prefer the controllable to the real.
This can quickly become a cycle that is hard to notice until it is entrenched. A pattern that I am noticing in more of in my practice as a psychologist are patients, mostly men, who come in dissatisfied with their dating lives, but do not always recognize the porn they are consuming as the culprit. The men can perform sexually, but struggle with emotional connection. They want partnership, but the negotiation and compromise of opening up in early dating feels exhausting. So, they start using interactive porn like webcam sites and live-streamed content and, without really noticing it, their use increases insidiously.
They are not consciously choosing to avoid dating; they say they want a relationship. However, over time, this more interactive porn becomes a central feature of their lives. Sometimes they come in worried about that habit, but more often, I am the one who has to point out that their porn use has eroded their ability to connect and their desire to try.
What I am seeing is not anecdotal. Research suggests that when people move from watching porn alone to using interactive content, they are more likely to struggle with intimacy and relationships. What is driving the challenge — users getting a feeling of connection without having to risk anything — is instructive. Combined with what has been found so far about AI’s effects on romantic relationships, it helps explain the healthy societal norms pornographic deepfakes can disrupt.
Every real relationship skill gets built through productive conflict — disappointment, compromise and communication — not through effortless, frictionless fantasy. If a person never has to subject themselves to someone saying no, stumble through explaining what they want, or suffer the indignity or embarrassment of things going wrong, they are not developing the capabilities of sustaining a real relationship.
We are quick to tell people that nonconsensual image generation is wrong because it violates the person being depicted. That part of the message is essential, but it is only half the story. We also need to constantly tell users that they would become less able to find satisfaction with real partners and ultimately lonelier. As AI rapidly changes and more impressionable young people get access to it, getting that fuller warning out might stop someone before they ever rationalize trying something this harmful.
Sarah Gundle is a psychologist in private practice and an assistant clinical professor at the Icahn School of Medicine at Mount Sinai Medical Center. This column reflects the personal views of the author and does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
The conflict in the Middle East has been disrupting financial markets, raising concerns about rising inflationary pressures and global economic growth. One market that some investors are particularly worried about has not been heavily covered in the news: the private credit market. Even before the joint US-Israeli attacks on Iran on Feb. 28, global capital markets had faced growing structural pressure — the deteriorating funding conditions in the private credit market. The private credit market is where companies borrow funds directly from nonbank financial institutions such as asset management companies, insurance companies and private lending platforms. Its popularity has risen since
The Donald Trump administration’s approach to China broadly, and to cross-Strait relations in particular, remains a conundrum. The 2025 US National Security Strategy prioritized the defense of Taiwan in a way that surprised some observers of the Trump administration: “Deterring a conflict over Taiwan, ideally by preserving military overmatch, is a priority.” Two months later, Taiwan went entirely unmentioned in the US National Defense Strategy, as did military overmatch vis-a-vis China, giving renewed cause for concern. How to interpret these varying statements remains an open question. In both documents, the Indo-Pacific is listed as a second priority behind homeland defense and
Every analyst watching Iran’s succession crisis is asking who would replace supreme leader Ayatollah Ali Khamenei. Yet, the real question is whether China has learned enough from the Persian Gulf to survive a war over Taiwan. Beijing purchases roughly 90 percent of Iran’s exported crude — some 1.61 million barrels per day last year — and holds a US$400 billion, 25-year cooperation agreement binding it to Tehran’s stability. However, this is not simply the story of a patron protecting an investment. China has spent years engineering a sanctions-evasion architecture that was never really about Iran — it was about Taiwan. The
In an op-ed published in Foreign Affairs on Tuesday, Chinese Nationalist Party (KMT) Chairwoman Cheng Li-wun (鄭麗文) said that Taiwan should not have to choose between aligning with Beijing or Washington, and advocated for cooperation with Beijing under the so-called “1992 consensus” as a form of “strategic ambiguity.” However, Cheng has either misunderstood the geopolitical reality and chosen appeasement, or is trying to fool an international audience with her doublespeak; nonetheless, it risks sending the wrong message to Taiwan’s democratic allies and partners. Cheng stressed that “Taiwan does not have to choose,” as while Beijing and Washington compete, Taiwan is strongest when
RSS