Tom Newhouse, vice president of Convergence Media, discusses the potential impact of artificial intelligence on elections after an RNC AI ad garnered attention. As Elon Musk and others continue to sound the alarm about the potential dangers of artificial intelligence, an unlikely duo of a data scientist and a political philosopher is teaming up to use AI with a different purpose in mind: bridging society's increasingly stark political divisions. The project stemmed from the research of David Rozado, a professor at Te Pūkenga -- the New Zealand Institute of Skills and Technology, who's recent work has drawn attention to political bias in ChatGPT and the potential for such bias in other AI systems. Rozado found that in 14 out of 15 political orientation test answers from ChatGPT, a product of the company OpenAI, were deemed as giving left-leaning viewpoints. At the same time, however, the AI language processing tool denied having any political bias or orientation, maintaining that it was just providing objective and accurate information to users. "The system would flag as hateful comments about certain groups but not others," Rozado told Fox News Digital, noting for example that the system would say it's hateful to call women dishonest but not men.

Next year's elections in Britain and the US could be marked by a wave of AI-powered disinformation, experts have warned, as generated images, text and deepfake videos go viral at the behest of swarms of AI-powered propaganda bots. Sam Altman, CEO of the ChatGPT creator, OpenAI, told a congressional hearing in Washington this week that the models behind the latest generation of AI technology could manipulate users. "The general ability of these models to manipulate and persuade, to provide one-on-one interactive disinformation is a significant area of concern," he said. "Regulation would be quite wise: people need to know if they're talking to an AI, or if content that they're looking at is generated or not. The ability to really model … to predict humans, I think is going to require a combination of companies doing the right thing, regulation and public education."

Fine-grained financial sentiment analysis on news headlines is a challenging task requiring human-annotated datasets to achieve high performance. Limited studies have tried to address the sentiment extraction task in a setting where multiple entities are present in a news headline. In an effort to further research in this area, we make publicly available SEntFiN 1.0, a human-annotated dataset of 10,753 news headlines with entity-sentiment annotations, of which 2,847 headlines contain multiple entities, often with conflicting sentiments. We augment our dataset with a database of over 1,000 financial entities and their various representations in news media amounting to over 5,000 phrases. We propose a framework that enables the extraction of entity-relevant sentiments using a feature-based approach rather than an expression-based approach. For sentiment extraction, we utilize 12 different learning schemes utilizing lexicon-based and pre-trained sentence representations and five classification approaches. Our experiments indicate that lexicon-based n-gram ensembles are above par with pre-trained word embedding schemes such as GloVe. Overall, RoBERTa and finBERT (domain-specific BERT) achieve the highest average accuracy of 94.29% and F1-score of 93.27%. Further, using over 210,000 entity-sentiment predictions, we validate the economic effect of sentiments on aggregate market movements over a long duration.

Despite much success in natural language processing (NLP), pre-trained language models typically lead to a high computational cost during inference. Multi-exit is a mainstream approach to address this issue by making a trade-off between efficiency and accuracy, where the saving of computation comes from an early exit. However, whether such saving from early-exiting is robust remains unknown. Motivated by this, we first show that directly adapting existing adversarial attack approaches targeting model accuracy cannot significantly reduce inference efficiency. To this end, we propose a simple yet effective attacking framework, SAME, a novel slowdown attack framework on multi-exit models, which is specially tailored to reduce the efficiency of the multi-exit models. By leveraging the multi-exit models' design characteristics, we utilize all internal predictions to guide the adversarial sample generation instead of merely considering the final prediction. Experiments on the GLUE benchmark show that SAME can effectively diminish the efficiency gain of various multi-exit models by 80% on average, convincingly validating its effectiveness and generalization ability.

We present an end-to-end diffusion-based method for editing videos with human language instructions, namely InstructVid2Vid. Our approach enables the editing of input videos based on natural language instructions without any per-example finetuning or inversion. The proposed InstructVid2Vid model combines a pretrained image generation model, Stable Diffusion, with a conditional 3D U-Net architecture to generate time-dependent sequence of video frames. To obtain the training data, we incorporate the knowledge and expertise of different models, including ChatGPT, BLIP, and Tune-a-Video, to synthesize video-instruction triplets, which is a more cost-efficient alternative to collecting data in real-world scenarios. To improve the consistency between adjacent frames of generated videos, we propose the Frame Difference Loss, which is incorporated during the training process. During inference, we extend the classifier-free guidance to text-video input to guide the generated results, making them more related to both the input video and instruction. Experiments demonstrate that InstructVid2Vid is able to generate high-quality, temporally coherent videos and perform diverse edits, including attribute editing, change of background, and style transfer. These results highlight the versatility and effectiveness of our proposed method. Code is released in InstructVid2Vid.

Ensembling BERT models often significantly improves accuracy, but at the cost of significantly more computation and memory footprint. In this work, we propose Multi-CLS BERT, a novel ensembling method for CLS-based prediction tasks that is almost as efficient as a single BERT model. Multi-CLS BERT uses multiple CLS tokens with a parameterization and objective that encourages their diversity. Thus instead of fine-tuning each BERT model in an ensemble (and running them all at test time), we need only fine-tune our single Multi-CLS BERT model (and run the one model at test time, ensembling just the multiple final CLS embeddings). To test its effectiveness, we build Multi-CLS BERT on top of a state-of-the-art pretraining method for BERT (Aroca-Ouellette and Rudzicz, 2020). In experiments on GLUE and SuperGLUE we show that our Multi-CLS BERT reliably improves both overall accuracy and confidence estimation. When only 100 training samples are available in GLUE, the Multi-CLS BERT_Base model can even outperform the corresponding BERT_Large model. We analyze the behavior of our Multi-CLS BERT, showing that it has many of the same characteristics and behavior as a typical BERT 5-way ensemble, but with nearly 4-times less computation and memory.

The robustness of recommender systems under node injection attacks has garnered significant attention. Recently, GraphRfi, a GNN-based recommender system, was proposed and shown to effectively mitigate the impact of injected fake users. However, we demonstrate that GraphRfi remains vulnerable to attacks due to the supervised nature of its fraudster detection component, where obtaining clean labels is challenging in practice. In particular, we propose a powerful poisoning attack, MetaC, against both GNN-based and MF-based recommender systems. Furthermore, we analyze why GraphRfi fails under such an attack. Then, based on our insights obtained from vulnerability analysis, we design an adaptive fraudster detection module that explicitly considers label uncertainty. This module can serve as a plug-in for different recommender systems, resulting in a robust framework named PDR. Comprehensive experiments show that our defense approach outperforms other benchmark methods under attacks. Overall, our research presents an effective framework for integrating fraudster detection into recommendation systems to achieve adversarial robustness.

This paper argues that Machine Learning (ML) algorithms must be educated. ML-trained algorithms' moral decisions are ubiquitous in human society. Sometimes reverting the societal advances governments, NGOs and civil society have achieved with great effort in the last decades or are yet on the path to be achieved. While their decisions have an incommensurable impact on human societies, these algorithms are within the least educated agents known (data incomplete, un-inclusive, or biased). ML algorithms are not something separate from our human idiosyncrasy but an enactment of our most implicit prejudices and biases. Some research is devoted to "responsibility assignment" as a strategy to tackle immoral AI behaviour. Yet this paper argues that the solution for AI ethical decision-making resides in algorithm education" (as opposed to the "training") of ML. Drawing from an analogy between ML and child education for social responsibility, the paper offers clear directions for responsible and sustainable AI design, specifically with respect to how to educate algorithms to decide ethically.

It's been a while since we had a truly great movie about devious, dystopian AIs priming themselves to take over the world, in which the key choices made by mere humans will decide whether we end up as just an organic footnote in histories written by our machine conquerors. Alex Garland's Ex-Machina (2014) springs to mind, while 2015's Avengers: Age of Ultron was a fun comic book romp, if lacking the spiky gravitas and sly intellectual thrust of Garland's debut. Now there's Gareth Edwards' The Creator, the first trailer for which debuted this week, arriving just as very real concerns about the ability of artificial intelligence to really muck things up for us humans are rearing their terrifying digital heads. At first glance, it looks as if Edwards has thrown in all our favourite sci-fi tropes. The basic scenario – tooled up military man fails in mission to wipe out robot child because she is just too cute – reminds us of kind-hearted Din Djarin's inability to bounty hunt Grogu in early episodes of The Mandalorian.

During my initial interactions with ChatGPT Plus, I was not fully convinced that OpenAI's $20-a-month subscription was worth it. While it was quite fun to test the upgraded chatbot powered by GPT-4, the free version seemed good enough for most prompts. I'm not a software developer who needs a deft coding assistant; I'm a nerd who uses chatbots to have entertaining conversations with artificial intelligence and brainstorm a little. On May 12, OpenAI announced that users who pay for ChatGPT Plus would be able to access beta versions of its chatbot with web browsing and plugins. Curious about the new features, I eschewed an evening of takeout, ate some gross leftovers, and spent money on finally upgrading my personal ChatGPT account.