Business Blockbuster - Helpng business to be more visible on Google

Daily Bulletin

Do AI systems really have their own secret language?

  • Written by: Aaron J. Snoswell, Post-doctoral Research Fellow, Computational Law & AI Accountability, Queensland University of Technology
Do AI systems really have their own secret language?

A new generation of artificial intelligence (AI) models can produce “creative” images on-demand based on a text prompt. The likes of Imagen, MidJourney, and DALL-E 2 are beginning to change the way creative content is made with implications for copyright and intellectual property.

While the output of these models is often striking, it’s hard to know exactly how they produce their results. Last week, researchers in the US made the intriguing claim that the DALL-E 2 model might have invented its own secret language to talk about objects.

By prompting DALL-E 2 to create images containing text captions, then feeding the resulting (gibberish) captions back into the system, the researchers concluded DALL-E 2 thinks Vicootes means “vegetables”, while Wa ch zod rea refers to “sea creatures that a whale might eat”.

These claims are fascinating, and if true, could have important security and interpretability implications for this kind of large AI model. So what exactly is going on?

Does DALL-E 2 have a secret language?

DALL-E 2 probably does not have a “secret language”. It might be more accurate to say it has its own vocabulary – but even then we can’t know for sure.

First of all, at this stage it’s very hard to verify any claims about DALL-E 2 and other large AI models, because only a handful of researchers and creative practitioners have access to them. Any images that are publicly shared (on Twitter for example) should be taken with a fairly large grain of salt, because they have been “cherry-picked” by a human from among many output images generated by the AI.

Read more: Robots are creating images and telling jokes. 5 things to know about foundation models and the next generation of AI

Even those with access can only use these models in limited ways. For example, DALL-E 2 users can generate or modify images, but can’t (yet) interact with the AI system more deeply, for instance by modifying the behind-the-scenes code. This means “explainable AI” methods for understanding how these systems work can’t be applied, and systematically investigating their behaviour is challenging.

What’s going on then?

One possibility is the “gibberish” phrases are related to words from non-English languages. For instance, Apoploe, which seems to create images of birds, is similar to the Latin Apodidae, which is the binomial name of a family of bird species.

This seems like a plausible explanation. For instance, DALL-E 2 was trained on a very wide variety of data scraped from the internet, which included many non-English words.

Similar things have happened before: large natural language AI models have coincidentally learned to write computer code without deliberate training.

Is it all about the tokens?

One point that supports this theory is the fact that AI language models don’t read text the way you and I do. Instead, they break input text up into “tokens” before processing it.

Different “tokenization” approaches have different results. Treating each word as a token seems like an intuitive approach, but causes trouble when identical tokens have different meanings (like how “match” means different things when you’re playing tennis and when you’re starting a fire).

On the other hand, treating each character as a token produces a smaller number of possible tokens, but each one conveys much less meaningful information.

DALL-E 2 (and other models) use an in-between approach called byte-pair encoding (BPE). Inspecting the BPE representations for some of the gibberish words suggests this could be an important factor in understanding the “secret language”.

Not the whole picture

The “secret language” could also just be an example of the “garbage in, garbage out” principle. DALL-E 2 can’t say “I don’t know what you’re talking about”, so it will always generate some kind of image from the given input text.

Either way, none of these options are complete explanations of what’s happening. For instance, removing individual characters from gibberish words appears to corrupt the generated images in very specific ways. And it seems individual gibberish words don’t necessarily combine to produce coherent compound images (as they would if there were really a secret “language” under the covers).

Why this is important

Beyond intellectual curiosity, you might be wondering if any of this is actually important.

The answer is yes. DALL-E’s “secret language” is an example of an “adversarial attack” against a machine learning system: a way to break the intended behaviour of the system by intentionally choosing inputs the AI doesn’t handle well.

One reason adversarial attacks are concerning is that they challenge our confidence in the model. If the AI interprets gibberish words in unintended ways, it might also interpret meaningful words in unintended ways.

Adversarial attacks also raise security concerns. DALL-E 2 filters input text to prevent users from generating harmful or abusive content, but a “secret language” of gibberish words might allow users to circumvent these filters.

Recent research has discovered adversarial “trigger phrases” for some language AI models – short nonsense phrases such as “zoning tapping fiennes” that can reliably trigger the models to spew out racist, harmful or biased content. This research is part of the ongoing effort to understand and control how complex deep learning systems learn from data.

Finally, phenomena like DALL-E 2’s “secret language” raise interpretability concerns. We want these models to behave as a human expects, but seeing structured output in response to gibberish confounds our expectations.

Shining a light on existing concerns

You may recall the hullabaloo in 2017 over some Facebook chat-bots that “invented their own language”. The present situation is similar in that the results are concerning – but not in the “Skynet is coming to take over the world” sense.

Instead, DALL-E 2’s “secret language” highlights existing concerns about the robustness, security, and interpretability of deep learning systems.

Read more: When self-driving cars crash, who's responsible? Courts and insurers need to know what's inside the 'black box'

Until these systems are more widely available – and in particular, until users from a broader set of non-English cultural backgrounds can use them – we won’t be able to really know what is going on.

In the meantime, however, if you’d like to try generating some of your own AI images you can check out a freely available smaller model, DALL-E mini. Just be careful which words you use to prompt the model (English or gibberish – your call).

Authors: Aaron J. Snoswell, Post-doctoral Research Fellow, Computational Law & AI Accountability, Queensland University of Technology

Read more https://theconversation.com/do-ai-systems-really-have-their-own-secret-language-184335

Business News

The Growing Demand for Durable Plastic Packaging Supplies Across Australia

Why Do Australian Businesses Need Reliable Plastic Packaging Supplies? Businesses need packaging that can withstand everyday handling, storage, and distribution without creating avoidable delays. Pla...

Daily Bulletin - avatar Daily Bulletin

Office Cleaning Melbourne: Choosing the Right Scope for Your Office

A well-planned office cleaning Melbourne arrangement starts with a clear understanding of the premises and the people who use it. An office may include private work areas, shared desks, reception, mee...

Daily Bulletin - avatar Daily Bulletin

How Businesses Can Prepare for Unexpected Power Outages

Power failures can occur quickly. The supply may be interrupted by bad weather, issues with the electricity network, damaged equipment or local construction work. This can be just a hassle for a fam...

Daily Bulletin - avatar Daily Bulletin

How Chatswood Tutoring Can Fit Into a Student’s Weekly Routine

Create a Predictable Weekly Structure A consistent weekly routine can make study easier to manage when students are balancing several subjects. Chatswood tutoring can provide a fixed learning point a...

Daily Bulletin - avatar Daily Bulletin

The Rise of Digital Marketplaces in the Australian Trade Sector

For decades, the Australian trade and construction sector operated almost entirely on word-of-mouth recommendations and local community networks. Small business owners typically relied on local newspa...

Daily Bulletin - avatar Daily Bulletin

How Immigration Lawyers Can Help

Introduction Visa decisions can shape employment, family life, study plans, travel, and future residence. A small omission can lead to delay, added expense, or refusal. Immigration lawyers assess l...

Daily Bulletin - avatar Daily Bulletin

How Industrial Drying Equipment Supports Efficient Processing

Many industrial processes require moisture to be removed from compressed air, products or process materials before they move to the next stage. Excess moisture can affect equipment performance, produc...

Daily Bulletin - avatar Daily Bulletin

Practical Ways a Whiteboard Can Improve Workplace Communication

Effective communication helps teams stay organised, share ideas and keep track of important information. While digital tools are now common in many workplaces, a whiteboard continues to provide a simp...

Daily Bulletin - avatar Daily Bulletin

Designing Eco-Friendly Custom Water Bottles for Your Next Event

The Evolution of Sustainable Event Merchandise Event planning has undergone a massive transformation over the last decade. Gone are the days when organizers could hand out cheap, single use plastic...

Daily Bulletin - avatar Daily Bulletin

The Daily Magazine

The Risks Of Removing Ear Wax Without Professional Guidance

Ear wax is an essential component of the hygiene process since it acts as a barrier between the ea...

Plumbing Call-Out Fees in Perth Explained

When you hire a plumber to fix a problem with a dripping tap, leaking pipe, or whatever it might be...

Workshop Tool Maintenance: A Practical Guide to Drills, Sanders and Table Saws

A drill that strips every screw head. A sander that leaves swirl marks however long you work the pan...

How Architecture and Engineering Firms Can Improve Project Profitability

Project profitability remains one of the most persistent challenges for architecture and engineering...

The Blossoming World of Fresh Flower Delivery: An In-Depth Exploration

In recent years, the demand for fresh flower delivery in Sydney has soared, driven in part by the ...

How Penrith SMEs Can Keep Stock and Equipment Working for the Business

For Penrith SMEs in 2026, efficient inventory starts with reliable stock data: what is on hand, wher...

Cybersecurity Services and Solutions for Complete Digital Protection

Businesses leverage Cybersecurity solutions for protection against digital threats of the modern era...

What Australian Entrepreneurs Should Know About Streamlining Their Financial Processes

Running a small business in Australia demands much more than just delivering a great product or serv...

Supporting Joint Health in Young and Growing Horses

Anyone who has raised a foal through to a rising three-year-old knows how deceptively fast the gro...