Three ways artificial intelligence (AI) is putting gender equality at risk.

Like many things that hold revolutionary potential, artificial intelligence (AI) is creating a unique and unprecedented set of opportunities and challenges.

As we unlock its capabilities, it is equally important to understand how Australians feel about the risks and whether they trust how AI is being developed and used. Because without trust, the benefits of AI will not be realised.

In an increasingly digital world, AI is developing at speed and scale, often outpacing the regulatory frameworks guiding its use and the safeguards needed to maintain public trust.

In fact, research commissioned by Minderoo Foundation found almost two-thirds of Australians believe AI is moving too fast and most say the risks outweigh the benefits.

The results of this national survey also uncovered an AI gender divide whereby women were much more likely to say the risks outweigh the benefits than men. Women were also more likely to say AI was moving too fast (66 per cent) compared with men (59 per cent).

And it’s not surprising. While AI has the potential to drive positive social progress, it also risks reinforcing existing gender inequalities and, in some cases, amplifying them. When people see harm without safeguards, trust erodes quickly, slowing adoption and increasing pressure for stricter regulation.

Perhaps most alarmingly, these flawed systems are increasingly shaping how young people understand gender, relationships and power.

Here are three key concerns around artificial intelligence and what it means for progress on gender equality.

Algorithms are the building blocks of artificial intelligence. On social media platforms, they act as pattern recognition systems, delivering content that captures our attention, aligns with our interests, and – for better or for worse – drives engagement.

But beyond creating entertainment value, these systems have a dark side that amplify harms.

Research shows that misogynistic content can be rapidly surfaced and reinforced, with harmful ideologies and tropes moving beyond screens and into school yards, becoming embedded in mainstream youth culture.

This is partly driven by platforms prioritising engagement over safety, creating feedback loops that reward more divisive content.

A recent study found a fourfold increase in misogynistic content suggested by TikTok over just five days, as the algorithm served progressively more extreme material, often centred on anger and blame directed at women.

A 2024 study by Dublin City University found that on platforms like TikTok and YouTube Shorts, all male-identified accounts were fed misogynistic or extremist material, irrespective of whether they sought it out, and the accounts all received this content within the first 23 minutes of signing up.

And we are already seeing the real-world consequences of this.

Findings from the 2026 International Women’s Day Report by King’s College London showed that Gen Z men are more likely to hold traditional views on gender roles than any other generation before them. Almost a third agree that a wife should obey her husband, a higher proportion than among Baby Boomers.

All of this brings into focus just how urgent the challenge is to strengthen accountability in the digital environment.

This means placing greater responsibility on platforms to protect users from harm and ensuring regulatory frameworks, such as the Online Safety Act, are fit for purpose so people can have confidence that AI systems are being managed responsibly. Clear, enforceable rules are critical to building that confidence.

At the same time, we must support the work already underway in Australia to respond to these challenges.

Digital Rights Watch continues to push for reforms that uphold privacy and data rights and strengthen accountability in the online space. Its recently launched AI Harms register is a timely call to action, seeking stories from Australians who have experienced harm from AI tools and want to help the campaign for more protections.

Teach Us Consent is advocating for stronger protections against technology-facilitated harm, including addressing the role of algorithms in amplifying harmful content and championing an opt-out feature.

The Human Technology Institute, based at the University of Technology Sydney, sits at the forefront of responsible technology policy, driving meaningful reform, capability uplift and action in the Australian AI regulatory space. The team has been working closely with company directors and senior leaders from government, industry and civil society to drive practical, human-centred AI governance best practice across both the Australian profit and not-for-profit sectors.

Additionally, approaching the issue from an alternative angle, Equimundo is working to foster healthier forms of masculinity, including addressing the role of online environments in shaping behaviours and attitudes. Their work focuses on interrupting pathways to harm and creating healthier spaces for young men to build positive connections.

Cumulatively, this represents a coordinated effort to respond to a rapidly evolving challenge. But unlocking the benefits of AI depends on public confidence that clear safeguards are in place.

Government has a unique role in setting enforceable standards and strengthening accountability. And the public agrees: 38 per cent of Australians think Government should be responsible for managing the risks of AI, whereas 27 per cent believe the responsibility sits with tech companies and AI providers.

From there, industry, philanthropy and civil society can build on that foundation to support prevention, education and long-term progress for gender equality.

But the pitfalls don’t end there…

A deepfake is an image or video in which a real person’s face or body is altered using AI software to create a false depiction.

AI now provides the tools for anyone to create highly realistic, manipulated imagery, including fake sexualised material.

And as the use of AI increases, so too does its misuse, including the creation of explicit deepfakes without consent.

This issue is not confined to anonymous corners of the internet.

Image-based abuse involving AI is becoming increasingly pervasive, including in schools.

Not only is this issue plaguing our youth, but it is also showing up as another tool used in the arsenal of abusers, particularly against those already at high risk of intimate parter violence and coercive control.

For many victim-survivors, this forms part of a broader pattern of abuse, as perpetrators employ diverse tactics to exert control, with the ever-evolving technological landscape providing new avenues to do so.

All of this reinforces the need for stronger safeguards to maintain trust and protect people from harm.

This includes limiting the creation and distribution of non-consensual content and strengthening accountability across platforms.

But regulation alone is not enough.

Although these tools exist and are accessible, it is still people choosing to use them in harmful and, in many cases, abusive ways.

Therefore, while strengthened laws preventing misuse will be critical, this response must sit alongside broader efforts focused on prevention, including education and early intervention, particularly with young people.

For parents looking for guidance in navigating online safety, check out eSafety Commissioner resources for schools and parents.

AI has the potential to transform many parts of our lives, but it also risks reinforcing existing biases and embedding them at scale. When bias is left unchecked, it undermines confidence in how these systems make decisions.

This shows up everywhere from hiring decisions and loan approvals to legal judgements and health care diagnoses.

AI gender bias occurs when AI models treat people differently based on their gender, reflecting the patterns and assumptions present in the data they are trained on.

Because AI is trained on existing data, it can replicate and scale existing biases if left unchecked.

To counter this, data must be actively selected and designed to reflect diverse social backgrounds and experiences, while removing historical bias and stereotypes around gender roles.

If we fail to do this, we risk mirroring and reinforcing existing social structures, many of which carry expectations that can be harmful to all genders.

At the same time, when used intentionally, AI also has significant potential to support progress in this space by using it to identify and address gendered inequalities and bias.

For example, AI has been used to uncover gender pay gaps in the workforce by analysing tools like Glassdoor, and to challenge long-standing credit score biases, as seen with companies like Zest AI.

Ultimately, addressing these risks will require a combination of advocacy, policy reform, research, and innovation, including new tools and interventions.

This is essential to ensure that the potential of artificial intelligence is harnessed, without recreating or amplifying the very inequalities it claims to have the power to overcome.

Our world is rapidly changing, and technological innovation is outpacing the structures and regulations designed to guide it.

While the risks are significant, so is the opportunity to act. But it will require sustained focus and collective effort to build trust and deliver safe, responsible outcomes at scale.

Meaningful progress on gender equality will not happen by default. It must be actively protected and intentionally designed in the face of new challenges and growing resistance.

The decisions we make now will determine whether artificial intelligence becomes a tool for progress, or a force that sets us back. If we get the safeguards right, we unlock the benefits. If we don’t, we lose trust and the opportunity that comes with it.