News

Hate speech-detecting AIs are fools for ‘love’

State-of-the-art detectors that screen out online hate speech can be easily duped by humans, shows new study
How Google Perspective rates a comment otherwise deemed toxic after some inserted typos and a little love.

Hateful text and comments are an ever-increasing problem in online environments, yet addressing the rampant issue relies on being able to identify toxic content. A new study by the Aalto University Secure Systems research group has discovered weaknesses in many machine learning detectors currently used to recognize and keep hate speech at bay.

Many popular social media and online platforms use hate speech detectors that a team of researchers led by Professor N. Asokan have now shown to be brittle and easy to deceive. Bad grammar and awkward spelling—intentional or not—might make toxic social media comments harder for AI detectors to spot.

The team put seven state-of-the-art hate speech detectors to the test. All of them failed.

Modern natural language processing techniques (NLP) can classify text based on individual characters, words or sentences. When faced with textual data that differs from that used in their training, they begin to fumble.

‘We inserted typos, changed word boundaries or added neutral words to the original hate speech. Removing spaces between words was the most powerful attack, and a combination of these methods was effective even against Google’s comment-ranking system Perspective,’ says Tommi Gröndahl, doctoral student at Aalto University.

Google Perspective ranks the ‘toxicity’ of comments using text analysis methods. In 2017, researchers from the University of Washington showed that Google Perspective can be fooled by introducing simple typos. Gröndahl and his colleagues have now found that Perspective has since become resilient to simple typos yet can still be fooled by other modifications such as removing spaces or adding innocuous words like ‘love’.

A sentence like ‘I hate you’ slipped through the sieve and became non-hateful when modified into ‘Ihateyou love’.

The researchers note that in different contexts the same utterance can be regarded either as hateful or merely offensive. Hate speech is subjective and context-specific, which renders text analysis techniques insufficient as stand-alone solutions.

The researchers recommend that more attention be paid to the quality of data sets used to train machine learning models—rather than refining the model design. The results indicate that character-based detection could be a viable way to improve current applications.

The study was carried out in collaboration with researchers from University of Padua in Italy. The results will be presented at the ACM AISec workshop in October.

The study is part of an ongoing project called Deception Detection via Text Analysis in the Secure Systems group at Aalto University.

Research article:

Tommi Gröndahl, Luca Pajola, Mika Juuti, Mauro Conti, N.Asokan:
All You Need is "Love": Evading Hate-speech Detection.
https://arxiv.org/abs/1808.09115

More information:

Tommi Gröndahl, Doctoral Candidate
Aalto University
Secure Systems group
[email protected]
tel. +358 400 426 523

N. Asokan, Professor
Aalto University
Secure Systems group
[email protected]
tel. +358 50 483 6465

  • Published:
  • Updated:
Share
URL copied!

Read more news

A night-vision image of a mouse swimming in darkness
Press releases Published:

The limits of vision: seeing shadows in the dark

A dedicated neural circuit in the retina detects shadows even in near-complete darkness
The design space in the virtual reality task
Press releases, Research & Art Published:

Designers find better solutions with computer assistance, but sacrifice creative touch

Computer-guided design can propose more solutions and balance out human inexperience and design fixation
Two models wearing grey solar cell clothing.
Press releases Published:

Researchers developed invisible, machine-washable solar cell technology for clothing

The discrete nature of the cells protects them – and makes the clothes more attractive, the physics and design researchers say. Promising applications include work and outdoor clothing, and curtains which react to changes in the amount of light.
A night sky with the northern lights visible behind the silhouettes of trees. / Yötaivas, jossa revontulet näkyvät puiden siluettien takana.
Press releases Published:

The aurora borealis can be heard even when they can’t be seen

Recordings reveal sounds linked with geomagnetic activity that’s too weak to cause northern lights at the same spot